Pancreatic cancer is a lethal disease. Empirical development of new drugs has not resulted in any meaningful improvement in survival. New strategies are urgently needed to combat this disease. We have developed and optimized a low-passage xenograft model that may permit an individualized approach to the treatment of patients with pancreatic cancer. Fresh pancreatic cancer tissues obtained at the time of surgical resection are implanted in nude mice. Xenografted tumors can be treated with anticancer agent to determine their in vivo activity. In previous studies we have mastered this model and have obtained a high take on rate and excellent in vivo results. We have also shown that the xenografted tumors closely resemble at the histology, gene mutation, and selected gene expression levels the biological features of the primary tumor from which they were generated. Sucessive passage in mice does not influence the resistance/susceptibility properties of the tumors to the drugs we have preliminarily tested and does not result in mayor changes in biological features. Because tumors can be indefinitely propagated in the mice the model is also very well suited to investigate biological markers that predict response and/or resistance to drugs. This translational application is to utilize the above mentioned preclinical model to individualize the development of new drugs for patients with pancreatic cancer. The hypothesis to be tested is that model-based selection of drugs for patient's treatment will result in better outcome than expected with random selection. We propose the following three Specific Aims: 1) to determine the activity of a series of novel anticancer agents against a set of xenografted tumors obtained from patients with resected pancreatic cancer;b) to conduct a phase II clinical trial in which patients whose tumor was xenografted in the mice will be treated at the time of progression with model-selected new agents and;3) to explore biological markers of response to treatment agents in tumor tissues. We expect this approach will validate the use of the low-passage xenograft model to predict susceptibility to a drug. If correct, this model will then be invaluable to discover biomarkers predicting drug response and as a screening model to select drug for clinical development in pancreatic cancer.